What is it about?
Explainable Artificial Intelligence aims at making the internal reasoning of decision-making algorithms interpretable by humans. However, XAI methods have been criticized for being designed only for people who already have deep knowledge of the matter and not for lay users. In this work, we tackle this problem using a visual approach for two well-known AI methods: Rules and Feature Importance.
Featured Image
Photo by Andrew Schultz on Unsplash
Why is it important?
Research in the area of Explainable AI is beginning to use Human-centered (HCI) approaches. This paper seeks to apply some principles of HCI to achieve more intuitive explanations for non-expert use.
Perspectives
I think exploring the potential of visualization is very important for the development of a more user-centered approach to explainable AI. In addition, designing and conducting a user study was of great value to this project.
Eleonora Cappuccio
Universita degli Studi di Pisa
Read the Original
This page is a summary of: Demo: an Interactive Visualization Combining Rule-Based and Feature Importance Explanations, September 2023, ACM (Association for Computing Machinery),
DOI: 10.1145/3605390.3610811.
You can read the full text:
Contributors
The following have contributed to this page
